摘要
针对红外图像的模糊性和非广延性,提出一种基于模糊Tsallis熵的快速阈值分割方法。先用模糊隶属度函数把图像灰度直方图转换到模糊域,分别定义图像背景与目标的模糊Tsallis熵,并根据不可加熵的伪叠加原理求出图像总熵。在参数寻优过程中,提出一种基于Logistic映射的混沌蛙跳优化算法,根据最大熵原理对模糊隶属度函数进行参数优化,进而得到图像的最佳分割阈值。与典型的阈值法进行对比试验,结果证实了该方法的有效性和鲁棒性。
In view of the fuzziness and non-extensiveness of infrared image,a fast threshold segmentation method based on fuzzy Tsallis entropy was proposed.Firstly,the histogram of image was transformed into fuzzy domain by fuzzy membership functions,then the fuzzy Tsallis entropies of background and object were defined,and thus the total entropy of image was obtained according to the pseudo additivity rule of non-extensiveness entropy.In the process of parameters optimization,a kind of chaos shuffled frog-leaping algorithm based on logistic mapping was proposed.The parameters of fuzzy member functions were optimized according to the maximum entropy principle,and then the optimal threshold of image was obtained.Compared with typical threshold segmentation methods by experiment,the presented method was verified to be efficient and robust.
出处
《安徽农业科学》
CAS
2012年第7期4393-4395,4398,共4页
Journal of Anhui Agricultural Sciences
关键词
红外目标分割
模糊Tsallis熵
混沌蛙跳算法
最大熵原则
Infrared object segmentation
Fuzzy Tsallis entropy
Chaos shuffled frog-leaping algorithm
Maximum entropy principle
